The SU2 code was extended to include radial basis function interpolation across meshes in simulations involving fluid-structure interactions. This was required because the fluid and structure meshes have different requirements (i.e. the fluid meshes are generally finer and unstructured while the structure meshes are coarser and structured). You can see the different between the blue and red mesh in the diagram above. This causes a mismatch in the nodes at the interface and a way must be devised to transfer information across them. The pressure and shear forces are transferred from the fluid nodes to the structure nodes as boundary conditions for the next structure iteration. In the other direction, the deformation is transferred to the fluid mesh for mesh deformation to get a more accurate solution in the next iteration.

The interpolation routines were successfully implemented and validated against cases with matching meshes. It improved robustness of the code and we were able to get previously unsolved cases to converge. The code is available at my GitHub. Much more information can be found in my thesis.

This project was done as part of my Master’s thesis in Imperial College London. Honestly, I chose it because the project title Spatial Interpolation in Fluid-Structure Interaction with Radial Basis Functions sounded terribly intimidating. I didn’t understand half of the words in the title and I felt that I would be getting my money’s worth by taking on such a project. In hind side, I really did learn a lot from it and the concepts were manageable when I got down into the details of it. My hard work was rewarded with a best thesis award at the end of my Master’s program.